Healthcare in the PICU May Be More Complicated Than We Thought—Who Knew?*
In this issue of Pediatric Critical Care Medicine, Shein et al (5) examine a multicenter, administrative database and describe the use of neuropharmacologic drugs and resources related to neuromorbidity in mechanically ventilated children with bronchiolitis in the United States between 2006 and 2015. The authors demonstrate significant increases in the use of multiple neuroactive medications over time, with the largest relative increases observed for antipsychotic drugs (0.3% of patients in 2006 to 1.6% in 2015), clonidine (1.8% to 18.3%), dexmedetomidine (3.7% to 46.8%), hydromorphone (0.4% to 7.4%), and propofol (3.3% to 11.5%). Use of fentanyl, ketamine, lorazepam, methadone, morphine, and neuromuscular blocking agents were also noted to rise over the same timeframe, though the increases were not as pronounced. Of the examined surrogates for neuromorbidity, significant increases were observed in utilization of neurology consultation, occupational therapy (OT), physical therapy (PT), swallow evaluations, and brain MRI over the study period.
In interpreting these results, it is important to recognize that the epidemiology of hospitalized patients with bronchiolitis has changed significantly in recent decades. An examination of the Kids’ Inpatient Database demonstrated that hospitalizations for bronchiolitis of unknown viral etiology increased between 1997 and 2012 (6). More frequent use of mechanical ventilation and longer length of stay were also noted over time among patients with RSV and unspecified bronchiolitis, while RSV-associated mortality decreased over time. Growing medical complexity among hospitalized children and the effects of palivizumab provide some explanation for these findings. Concomitantly, management of bronchiolitis has continued to evolve; the use of noninvasive respiratory support, particularly high-flow nasal cannula, has been increasing, and considerable variability in the use of invasive ventilation occurs across centers (7, 8).
It is expected that these epidemiologic shifts factor into the findings of the present study, though the precise influence is unclear. Use of mechanical ventilation and neuromuscular blockade are inclusion criteria, obscuring patient-level characteristics that might be associated with changing requirements for these modalities over time. Though nearly half of the cohort from the present study had a significant comorbidity at time of admission, administrative data limit acuity adjustment. Although the authors do their best to address this limitation by incorporating vasopressor use into multivariable models, the approach falls admittedly short of capturing clinical circumstances relevant to the use of sedative-analgesic medications and markers of neurologic morbidity.
The number of neuroactive drugs used per patient and the number of drug days per day of mechanical ventilation also increased over the study period. Again, the relation of such findings to patient acuity and length of stay is unclear. Growing emphases on quality improvement metrics provides one putative explanation of these trends.